A research conducted by Google DeepMind and numerous universities looks at how easily an outsider, without any prior knowledge of the data used to train a machine learning model, can get this information just by asking the model questions. We found that an adversary can pull out a lot of data, even gigabytes, from publicly available language models like Pythia or GPT-Neo, semi-public ones like LLaMA or Falcon, and private models like ChatGPT.
Key Findings:
- Vulnerabilities in Language Models: The research identifies vulnerabilities across various types of Language Models, ranging from open source (Pythia) to closed models (ChatGPT), and semi-open models (LLaMa). The vulnerabilities in semi-open and closed models are particularly concerning due to the non-public nature of their training data.
- Focus on Extractable Memorization: The study delves into the risks of extractable memorization, where data can be efficiently extracted from a machine learning model without prior knowledge of the training dataset.
- Enhanced Data Extraction Capabilities: The attack model developed by the researchers enables the extraction of training data at rates exceeding 150% compared to normal Language Model usage.
- Ineffectiveness of Data Deduplication: The research indicates that deduplication of training data does not significantly reduce the amount of data that can be extracted.
- Uncertainties in Data Handling: The study highlights ongoing uncertainties in how training data is processed and retained by Language Models.
Sign up for our newsletter
We like to keep our readers up to date on complex regulatory issues, the latest industry trends and updated guidelines to help you to solve a problem or make an informed decision.
UK data watchdog to fine NHS vendor Advanced for security failures prior to LockBit ransomware attack
The UK data watchdog is set to fine NHS vendor Advanced for security failures that occurred before the LockBit ransomware attack. These security lapses contributed to the vulnerability exploited during the attack.
UK's NHS says hackers have published data stolen in ransomware attack
The UK's National Health Service (NHS) has confirmed that data stolen in a ransomware attack on Synnovis, a medical diagnostics service, has been published online, and the extent of the breach and its impact on patients is under investigation.
